BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
X-LIC-LOCATION:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:19911015T033000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251218T030656Z
LOCATION:Meeting Room S426+S427\, Level 4
DTSTART;TZID=Asia/Hong_Kong:20251215T151100
DTEND;TZID=Asia/Hong_Kong:20251215T152200
UID:siggraphasia_SIGGRAPH Asia 2025_sess110_papers_1568@linklings.com
SUMMARY:DiffTex: Differentiable Texturing for Architectural Proxy Models
DESCRIPTION:Weidan Xiong, Yongli Wu, and Bochuan Zeng (Shenzhen University
 ); Jianwei Guo (Beijing Normal University); Dani Lischinski (The Hebrew Un
 iversity of Jerusalem); Daniel Cohen-Or (Tel Aviv University, Shenzhen Uni
 versity); and Hui Huang (Shenzhen University)\n\nSimplified proxy models a
 re commonly used to represent detailed architectural structures, reducing 
 storage requirements and enabling real-time rendering. However, the geomet
 ric simplifications inherent in proxies result in a loss of fine detail, m
 aking it essential for textures to compensate for this loss. Preserving th
 e rich texture information from the original dense architectural reconstru
 ctions remains a daunting task, particularly when working with unordered R
 GB photographs. We propose an automated method for generating realistic te
 xture maps for architectural proxy models at the texel level from an unord
 ered collection of registered photographs. Our approach establishes corres
 pondences between texels on a UV map and pixels in the input images, with 
 each texel’s color computed as a weighted blend of associated pixel values
 . Using differentiable rendering, we optimize blending parameters to ensur
 e photometric and perspective consistency, while maintaining seamless text
 ure coherence. Experimental results demonstrate the effectiveness and robu
 stness of our method across diverse architectural models and varying photo
 graphic conditions, enabling the creation of high-quality textures that pr
 eserve visual fidelity and structural detail.\n\nRegistration Category: Fu
 ll Access, Full Access Supporter\n\nSession Chair: Beibei Wang (Nanjing Un
 iversity)\n\n
END:VEVENT
END:VCALENDAR